pytorch/xla

Pull Requests Commits

moving torch_xla2 experimental code in-tree (#6463)

qihqi committed 1 year ago

Verified 4814d64c

Support multiple StableHLO Composite outputs (#6295)

chunnienc committed 1 year ago

Verified 535d398b

Fix global_device_count(), local_device_count() for single process on CUDA (#6022)

vanbasten23 committed 1 year ago

Verified 8fc8d57a

Reenable TestPoolingNNDeviceTypeXLA (#6453)

qihqi committed 1 year ago

Verified 3e684095

Force guards when performing boolean operations on XLASymNode (#6433)

ezyang committed 1 year ago

Verified fe604e97

[benchmarks] Add support for AMP execution. (#6447)

ysiraichi committed 1 year ago

Verified ad822bf0

torchbench_model.py: fix YAML skip file path (#6457)

cota committed 1 year ago

Verified 510bb8ea

Unify TORCH_LIBRARY definitions (#6455)

alanwaketan committed 1 year ago

Verified 19923bd1

Disable test_EmbeddingBag_per_sample_weights_and_new_offsets (#6456)

JackCaoG committed 1 year ago

Verified 21653dc9

Fix benchmark process terminaion issue (#6441)

zpcore committed 1 year ago

Verified d1a454fe

Disable TestPoolingNNDeviceTypeXLA temporarily. (#6452)

yeounoh committed 1 year ago

Verified d0a5c34e

follow-up for #6420 (#6450)

ManfeiBai committed 1 year ago

Verified 89f7f43b

Add `benchmarks/dynamo` search option for finding PyTorch skip file. (#6446)

ysiraichi committed 1 year ago

Verified f4849154

Re-instate the missing CPU CI tests and their coverage reports (#6448)

yeounoh committed 1 year ago

Verified 52ca01e7

nightly.sh: also run with non-dynamo (i.e. lazy tensor) XLA (#6432)

cota committed 1 year ago

Verified c154b991

aggregate.py: fix plotting of more than 3 backends (#6449)

cota committed 1 year ago

Verified 800a3254

Enable install `google-cloud-sdk-gke-gcloud-auth-plugin` to enable TPUCI run_e2e_tests (#6420)

ManfeiBai committed 1 year ago

Verified 479559c4

Remove `TPU_C_API` device type (#6435)

will-cromar committed 1 year ago

Verified d686d65b

Properly lower aten::lerp (#6339)

wonjoo-wj committed 1 year ago

Verified d73bd833

Add some helpful debugging notes to TROUBLESHOOTING docs (#6430)

changm committed 1 year ago

Verified 17161eb6

Re-land "aggregate.py: add tab format (#6405)", reverted in #6424 (#6431)

cota committed 1 year ago

Verified 904676be

Move tests that need CUDA to `test_operations.py`. (#6160)

ysiraichi committed 1 year ago

Verified 16b6bd1b

Enable CUDA support for PyTorch on CI. (#6070)

ysiraichi committed 1 year ago

Verified d5e1b49e

Check precision only if on CUDA and not inductor. (#6407)

ysiraichi committed 1 year ago

Verified 30591ad8

Fix model skip yaml file not found issue (#6428)

zpcore committed 1 year ago

Verified e2012881

Revert "Disambiguate `benchmarks` directory search. (#6434)" (#6440)

zpcore committed 1 year ago

Verified bc4fee20

Disambiguate `benchmarks` directory search. (#6434)

ysiraichi committed 1 year ago

Verified 599b9256

update README.md doc to 2.2 (#6437)

zpcore committed 1 year ago

Verified ea9c6481

Avoid fallback for avg_pool - (#6409)

qihqi committed 1 year ago

Verified a0bae821

Fix distributed checkpoint errors when PyTorch CUDA is enabled. (#6421)

ysiraichi committed 1 year ago

Verified 85c66f5f

Newer Older